I have a question on subnetting. I've done alot of practice with class C, but realized I have not gone over anything in class b or a, while trying to create my own networks on packet tracer. I think i've figured it out, but would like some feedback/help provided I have done it wrong.
I have randomly given myself a Private IP of 172.29.143.0.
I have a subnet that needs 520 hosts, one that needs 200, one that needs 30, and then 4 subnets for router connections. This is what I've come up with:
500 hosts get IP range: 172.29.143.1-172.29.144.254 (with .0 for net and .255 for broadcast)
200 hosts get IP range: 172.29.145.1-172.29.145.254 (with .0 for net and .255 for broadcast)
30 hosts get IP range: 172.29.146.1-172.29.146.31 (with .0 for net and .32 for broadcast)
Each Router only needs 2 hosts, so the ranges are 172.29.146.34-.35; .37-.38; .40-.41; .43.44, with net and broadcasts inbetween.
I hope I did that right, if not I would appreciate any help you can give me. Thanks!
172.29.143.0 is not a valid Class B subnet. This IP is a valid host IP with all the class b subnets. You would need to assign yourself a supernet of 172.29.0.0/22 and then subnet it in order to create your networks.
So you would have the following.
500 hosts: 172.29.0.1 - .1.254 /23 (with .0 for net and .255 for broadcast)
200 hosts: 172.29.2.1 - .2.254 /24 (with .0 for net and .255 for broadcast)
30 hosts: 172.29.3.1 - .3.30 /28 (with .0 for net and .31 for broadcast)
The rest of your subnets would be /30 with the following ranges.
172.29.3.33 - .3.34 (with .32 for net and .35 for broadcast)
172.29.3.37 - .3.38 (with .36 for net and .39 for broadcast)
172.29.3.41 - .3.42 (with .40 for net and .43 for broadcast)
172.29.3.45 - .3.46 (with .44 for net and .47 for broadcast)
First of all, that clears up alot. I miswrote the first part, though. I actually wrote down 520 hosts on my network drawing, but only posted 500. But that also brings up another ? that I've not been able to answer. Do you always have to start out with X.0.0.0 for class A, X.X.0.0 for class B, and X.X.X.0 for class C?
When subnetting any network you have to think about powers of 2. 2 is an even number so when you subnet, you will never ever have a subnet start with an odd number unless you have a 255 in that octet.
In this case if you are going to use 172.29.x.x and your first subnet requires 500 hosts, you have to take a look at all of the subnets that are in 172.29.x.x that give you 500 hosts.
So with 500 hosts we need a /23 or 255.255.254.0, which gives us a network increment of 2 in the 3rd octet. With that we have subnets of 172.29.0.0 /23, 172.29.2.0 /23, all the way up to 172.29.254.0 /23
Now lets take a look at your random IP of 172.29.143.0. If this is a subnet of class B network 172.29.0.0, then you cannot go backwards and try to get 500 hosts, because this subnet would have a /24 or 255.255.255.0
If this is an IP address, then its not possible to tell what subnet it falls in without a subnet mask.
If you want to test yourself, start off by using a network address such as 172.29.0.0 /16, and then give yourself random requirements for hosts and/or subnets. Then try varying subnet masks, and try to subnet further.
BTW your math looks pretty good.
A class A reserves 8 network bits and uses 24 host bits
A class B reserves 16 network bits and uses 16 host bits
A class C reserves 24 network bits and uses 8 host bits.
When subnetting a Class A, B or C you are borrowing bits from the host bits to create a subnet.
So yes, a Class A always starts out as X.0.0.0, B as X.X.0.0 and C as X.X.X.0
Somewhat correct. The "default" masks are always as you said, a Class A starts out as X.0.0.0, B as X.X.0.0 and C as X.X.X.0 . However, I could be assigned the following Class B address from IANA:
Now when I subnet this network block for my company use I will start with the following as my mask 255.255.248.0 as this represents the block of IP addresses I was assigned by IANA.
22.214.171.124 - 126.96.36.199 mask 255.255.248.0 or /21
So, I can subnet this /21 block into 8 /24s as follows:
I can further subnet say the last /24 into a bunch of /30 for use on all my WAN links.
" " " "
" " " "
You get the idea. So, to answer his question,
"Do you always have to start out with X.0.0.0 for class A, X.X.0.0 for class B, and X.X.X.0 for class C?"
Answer: No, it depends on if you were given a starting mask or you assume the class "default".
Carve out your larger networks first. And a /16 gives you lots of different possibilities or subnetting! Decide what you need and how many of each.
And you may have "gaps" to allow for future expansion as well. It's just a planning exercise.
It's all about moving your line though. In RFC950's terms (e.g. history) once you subnetted, everyone had to have the same size mask. Not so much any longer. With VLSM, you can split things up to look like what Brian laid out.
This is my first post here i just found out about this site earlier.
I don't know how deep your subnetting skills are, but while i was studying for my CCNA my classmates had the same problem switching from subnetting Class C adresses into subnetting Class A and B.
Perhaps this will help you, or confuse you , but most people who had problems switching from class C -> A and B worked it out after learning that "you should always do the calculations in the octet that you are working in". In class C it's the last octet so they weren't confused, where as when you move it up to Class A and B you may be working in the 2nd and 3rd octet as well as the 4th depending on how you want to subnet.
To clarify: "doing the calculations in the octet that you are working in" simply means the octet in where you separate the subnetmask bits from the host bits in binary. Wheather it will be in the 2nd, 3rd or 4th octet doesn't matter - that's where you should do the increments for the subnets.
And let me give you a small advice for the CCNA-exam . Also practice doing supernetting and subnetting from a supernet. That will help you on your journey.
For example, you may get subnetting questions where the specified ip-network is a supernet (i.e 192.168.0.0 /22) and where all the IP-adresses belongs to a subnet from that range.
That being said i want to respectfully disagree with you Angelo. I agree that 172.29.0.0 /22 is a subnet of 172.29.0.0 / 16, but I disagree that it's a supernet because the subnetmask is less than or equal to then the classfull subnetmask of 172.29.0.0 / 16. I will claim that 172.29.0.0 /22 is a route summarize of several 172.29.x.x /24, but not a supernet. A supernet is route summarize of several Clasfull adresses.
I think you meant the process "supernetting", even so i wanted to clarify that it wasn't a supernet.
Edit: Oh i forgot to answer your initial question. (i assumed 520 hosts required in first network)
Start with the classfull adress: 172.29.0.0
Then subnet it from the largets network to the smallest.
172.29.0.0 / 22 - 520 hosts ( alot of adresses wasted/reserved )
172.29.4.0 / 24 - 200 hosts
172.29.5.0 / 27 - 30 hosts (a /28 mask like angelo suggested will only give you 14 host-adresses and not fit your requirements)
172.29.5.32 / 30 - 2 hosts (4-ip adresses)
172.29.5.36 / 30 - 2 hosts (4-ip adresses)
...and so on depending how many WAN-links you need.
In the first subnet you work in the 3rd octet, so you increment there (2 host bits)
In the second subnet you work in the 3rd octet, so you increment there (you use the whole 3rd octect for network bits, so next subnet starts at x.x.5)
In the third subnet you work in the 4th octet, so you increment there (5 host bits)
In the 4th and other WAN-subnets you wirk in the 4th octec, so you increment there (2 host bits).
....the same procedure works well for a class A network as well. There are exceptions but if you know your binary you should figure it out when it happens .
In regards to your post #9, "That being said i want to respectfully disagree with you Angelo. I agree that 172.29.0.0 /22 is a subnet of 172.29.0.0 / 16, but I disagree that it's a supernet because the subnetmask is less than or equal to then the classfull subnetmask of 172.29.0.0 / 16. I will claim that 172.29.0.0 /21 is a route summarize of several 172.29.x.x /24, but not a supernet. A supernet is route summarize of several Clasfull adresses."
When you write IP address blocks using the "slash" notation you are talking CIDR blocks rather then "classful" (A, B, C networks). So, if we start with a Class B (classful) address of 172.29.0.0 and subnet using the following mask 255.255.248.0, we get the following subnets:
172.29.0.0 <--- this is indeed a subnet of the larger "Classful" network
You are correct in that the 172.29.0.0 subnet mask 255.255.248.0 is a subnet of the larger "Classful" IP address block of 172.16.0.0. However, in terms of Classless Inter-Domain Routing (CIDR), it is a "supernet" of the smaller block of IP addresses or individual /24s. For example, the following are CIDR IP address blocks:
Since these all share the same CIDR routing prefix or common first 24 bits we can combine them into a single larger block of IP addresses with a new routing prefix. When we combine these address we are forming a supernet with a new "smaller" prefix length (21 < 24). This process is also called supernetting, route aggregation, or route summarization.
Here is a definition for supernet, "a supernet is a block of contiguous subnetworks addressed as a single subnet in the larger network. Supernets always have subnet masks that are "smaller"than the masks of the component networks.
Therefore, in terms of CIDR and the above definition, the IP address block 172.29.0.0/21 is a supernet of the contiguous individual /24 address blocks. Supernetting is not a summarization of "classful" addresses, but rather the combination of two or more networks or subnets with a common CIDR routing prefix.
Message was edited by: Brian
Hi Brian and thanks for your input, (and excuse my bad grammar)
In regards to your post above (#10):
Therefore, in terms of CIDR and the above definition, the IP address block 172.29.0.0/22 is a supernet of the contiguous individual /24 address blocks. Supernetting is not a summarization of "classful" addresses, but rather the combination of two or more networks or subnets with a common CIDR routing prefix."
I still disagree with that definition because i used to believe it was exactly like the way you are describing. That's how i learned about CIDR and prefixes and masks in the first place.
However i "only" got a 98.2 score on my final exam of CCNA2 because i answered wrong on a CIDR-question about supernets (not supernetting).
The question was exactly like the example you are describing above, it was a 172.x.x.x /22 summarizing a few 172.x.x.x /24 networks into a single route. And the question asked what you called the new prefix (the CIDR /22 prefix).
I answered a supernet and it was wrong. The correct answer should've been the Route summarize, route aggregation (and something else i can't remember). I was frustrated because of that question so I looked it up when i got back home from the exam.
And at least according to all CCNA-books from cisco-press I've read (CCNA,CCNP,Security,CCDA), your definition of a supernet:
"Here is a definition for supernet, "a supernet is a block of contiguous subnetworks addressed as a single subnet in the larger network. Supernets always have subnet masks that are "smaller"than the masks of the component networks."
Is in fact the definition of a route summarize, route aggregation. I checked this again before going to school, and that's how the CCNA-books define route-summarization and aggregation. When i get back from school i'll update with the quote from those books for reference .
I checked this up myself very thoroughly because i can't believe i failed that question...and I was wrong at the Exam and learned it the hard way. According to the cisco-books the 172.29.0.0 /16 network doesn't become a supernet until you make a supernet of it by giving it a subnetmask that is "smaller" (fewer network bits) than the classfull subnetmask (/15, /14 and so on).
"Supernetting is not a summarization of "classful" addresses, but rather the combination of two or more networks or subnets with a common CIDR routing prefix."
I agree that supernetting is not a summarization of classfull adresses. But a supernet is a summarization of classfull addresses into one route.
That's how cisco defines it in their curriculums and I got points reducted on my exam for mixing those up . If you have access to the books (CCNA/NP/DA,Security) check the glossary in the back of it and you'll see how Cisco defines it.
I don't like posting without accurate information so i'll quote it when i get home from school!
You are free to disagree, but you would be wrong. Supernetting, route summarization and route aggregation are terms used in CIDR and do not deal with "classful" network at all. Classful network already had "automatic" route summarization at the "classful" bit boundies. The creatation of Classless Inter-domain Routing (CIDR) replaced the old "Classful" addressing architecture.
Regarding the question you missed, the answer they give is wrong and you were correct. Again, the process of forming a supernet, which is often called "supernetting", is the same thing as route summarization or route aggregation.
So go with your original gut feeling. Supernetting is not the a summarization of "classful" networks, but a summarization of CIDR blocks. Supernetting requires the use of routing protocols that support classless inter-domain routing or CIDR. CIDR is based on VLSM. See RFC 950.
Here is another good example describing supernetting.
"CIDR provides the possibility of fine-grained routing prefix aggregation, also known as supernetting or route summarization. For example, sixteen contiguous /24 networks can be aggregated and advertised to a larger network as a single /20 route, if the first 20 bits of their network addresses match. Two aligned contiguous /20s may then be aggregated to a /19, and so forth. This allows a significant reduction in the number of routes that have to be advertised."
So, in the example posted in the discussion 172.29.0.0/22 is a supernet of four contiguous /24 routes. It is also a subnet of the "Classful" network 172.29.0.0/16. My post #10 mention a /22 but the IP address examples were referring to a /21. A /21 is eight contiguous /24 and a /22 is four contiguous /24. But still a supernet no the less.
Well this makes it interesting, because then i would've had 100% and the books need to be updates .
"You are free to disagree, but you would be wrong."
The only reason I disagree is because I got point reduction on the final exam, and when i looked it up in the books the definitions of a supernet and supernetting confirms the answer they wantes as "correct" on their test...and my gut feeling answer as wrong.
I do know of RFC 950, CIDR and VLSM. That's why my gut feeling on the test pushed me in that direction .
And like i've said above, in the cisco-books there is a difference between the process of supernetting and what a supernet is....please feel free to read their definitions when i get back from school, because believe me I would make the same "error?" again on a test.
I would love to straighten it out, since even today after so long time i still remember this issue because it "saved" me from getting a 100% score.
I am not sure which books you are referring, but the ones I have read support the definitions of supernetting. Can you give me the references that say supernetting is for "classful" networks. Very interested to see these.
I would not worry to much about the 100% vs the 98%. You were right in picking your answer, just because the test says otherwise, just means there are errors in the test.